ORIGINAL ARTICLE

cDNA-amplified fragment length polymorphism to studythe transcriptional responses of Lactobacillus rhamnosusgrowing in cheese-like mediumC.G. Bove, C. Lazzi, V. Bernini, B. Bottari, E. Neviani and M. Gatti

Department of Genetics, Biology of Microorganisms, Anthropology and Evolution, University of Parma, Parma, Italy

Introduction

Lactic acid bacteria (LAB) play different roles in cheese

making. Starter LAB (SLAB) participate in the fermenta-

tion process, whereas nonstarter LAB (NSLAB), probably

present in all cheeses, are involved in the maturation,

influencing flavour development (Beresford et al. 2001;

Beresford and Williams 2004; Broadbent and Steele 2005).

Although the role of NSLAB in cheese ripening has not

been clarified yet, different authors have suggested their

importance in this production step (Settanni and Mosch-

etti 2010). In Parmigiano Reggiano cheese, it was demon-

strated that SLAB are dominant until the second month

of ripening, and after cheese brining, the species NSLAB,

especially Lactobacillus rhamnosus, can grow and increase

in number, while SLAB cells undergo autolysis (De Dea

Lindner et al. 2008; Gatti et al. 2008). Lact. rhamnosus

was shown to be the dominant species present when

essential nutrients, such as sugars, are lacking. Therefore,

this species seems to well adapt to the absence of lac-

tose in cheese, confirming an optimal adaptability to

unfavourable growth conditions. Presumably, this charac-

teristic is because of the ability of Lact. rhamnosus to use

nitrogen fraction as an alternative energy source. Banks

and Williams (2004) reported that a potential energy

sources in maturing cheese include amino acids generated

during ripening. Peptides and amino acids can be

catabolized by lactobacilli via transamination reaction

(Kieronczyk et al. 2001; Yvon and Rijnen 2001; Liu et al.

2003). So far, only few studies have been carried out on

Lact. rhamnosus in cheese ripening (De Dea Lindner et al.

2008; Gatti et al. 2008; Neviani et al. 2009). The latest

Keywords

cDNA-AFLP, cheese-like conditions, gene

expression, Lactobacillus rhamnosus,

transcription profiling.

Correspondence

Claudio G. Bove, Department of Genetics,

Biology of Microorganisms, Anthropology and

Evolution, University of Parma, Viale Usberti

11 ⁄ A, 43100 Parma, Italy.

E-mail: claudiogiorgio.bove@nemo.unipr.it

2011 ⁄ 0271: received 16 February 2011,

revised 30 May 2011, accepted 27 June 2011

doi:10.1111/j.1365-2672.2011.05101.x

Abstract

Aims: Lactobacillus rhamnosus is a dominant species during Parmigiano

Reggiano cheese ripening and exhibits a great adaptability to unfavourable

growth conditions. Gene expression of a Lact. rhamnosus, isolated from

Parmigiano Reggiano cheese, grown in a rich medium (MRS) and in a cheese-

like medium (CB) has been compared by a novel cDNA-amplified fragment

length polymorphism (cDNA-AFLP) protocol.

Methods and Results: Two techniques, capillary and gel electrophoresis cDNA-

AFLP, were applied to generate unique transcript tags from reverse-transcribed

messenger RNA using the immobilization of biotinylated 3¢-terminal cDNA

fragments on streptavidin-coated Dynabeads. The use of three pairs of primers

allowed detecting 64 genes expressed in MRS and 96 in CB. Different tran-

scripts were observed when Lact. rhamnosus was cultured on CB and MRS.

Conclusions: The cDNA-AFLP approach proved to be able to show that

Lact. rhamnosus modifies the expression of a large part of genes when

cultivated in CB compared with growth under optimal conditions (MRS). In

particular, the profiles of the strain grown in CB were more complex probably

because the cells activate different metabolic pathways to generate energy and

to respond to the environmental changes.

Significance and Impact of Study: This is the first research on Lact. rhamnosus

isolated from cheese and represents one of the few concerning bacterial tran-

scriptomic analysis towards cDNA-AFLP approaches.

Journal of Applied Microbiology ISSN 1364-5072

ª 2011 The Authors

Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology 855

study demonstrated that the intraspecies heterogeneity,

found for 66 Lact. rhamnosus strains isolated during the

same Parmigiano Reggiano cheese making, seems to be

correlated with their abilities to adapt to the hostile envi-

ronment of the cheese throughout the ripening (Bove

et al. 2011). It was also found that the great part of

Lact. rhamnosus strains belonged to few biotypes that

were present in cheese from the first or second month of

ripening up to 10 months and beyond (Bove et al. 2011).

The detection of biotypes correlated with specific steps in

cheese ripening suggested that these strains may have spe-

cific metabolic activities, which allow an adaptation to

the microenvironment of cheese ripening.

The high degree of adaptation of Lact. rhamnosus to

different environments and conditions has been studied

towards several approaches (Succi et al. 2005; Koskenni-

emi et al. 2009) and probably involves many metabolic

and physiological changes. To identify and analyse the

genes involved in biological processes, genome-wide

expression analysis is considered one of the most efficient

tool (Breyne and Zabeau 2001). In this regard, cDNA-

amplified fragment length polymorphism (cDNA-AFLP)

is a method for genome-wide expression analysis often

proposed as an alternative to microarrays methods. In

fact, with respect to microarrays, at lower cost, cDNA-

AFLP has the advantage that any unknown genome or set

of genomes can be studied without prior sequence knowl-

edge and offers the possibility to detect also species-

specific genes that could be lacking in the reference strain.

Further, the cDNA-AFLP approach allows the detection

of lowly expressed genes and may allow the discrimina-

tion between homologous genes. Above all, the applica-

tion of cDNA-AFLP is suggested to study organisms with

complete genome sequence or cDNA collection unavail-

ability (Breyne and Zabeau 2001).

cDNA-AFLP is carried out according to the principle

of AFLP. The AFLP technique is based on the digestion

of DNA templates followed by the ligation of adapters to

restriction fragments and the selective PCR amplification

of subsets of these fragments using selective AFLP primers

containing a variable number of selective nucleotides.

Similarly, the original cDNA-AFLP method (Bachem et al.

1996) involves the reverse transcription of mRNA into

double-stranded cDNA, followed by the generation of a

complex mixture of transcript-derived fragments (TDFs)

by restriction enzyme digestion and ligation of specific

adapters, selective PCR amplification and, finally, the

visualization of the TDFs on high-resolution (sequence)

gels (Vuylsteke et al. 2007).

Recently, different researches were published to

improve the original cDNA-AFLP method (Decorosi et al.

2005; Kadota et al. 2007; Weiberg et al. 2008; Xiaohu

et al. 2009). Nevertheless, the most important variant is

the modification of the original protocol based on ‘one-

gene-multiple-tag’ into ‘one-gene-one-tag’ as reported by

Vuylsteke et al. (2007) and Breyne and Zabeau (2001).

The ‘one-gene-one-tag’ cDNA-AFLP protocol involves a

reduction in TDFs to a single fragment for each cDNA by

selecting the 3¢-terminal restriction fragment of each tran-

script before selective amplification. Recently, Levterova

et al. (2010) have developed a cDNA-AFLP strategy to

study the gene transcription in fungi based on the use of

capillary electrophoresis.

Although cDNA-AFLP technique has been extensively

applied to analyse transcription profiles of eukaryotic

organisms (Reijans et al. 2003; Bensch and Akesson 2005;

Ekkapongpisit et al. 2007; Neveu et al. 2007), there are

few data regarding cDNA-AFLP application to bacteria

(Dellagi et al. 2000; Noel et al. 2001). This is probably

due to the difficulties of working with bacterial mRNA. A

major technical challenge for transcriptome sequencing is

the low relative abundance of mRNAs in total cellular

RNA (1-5%) (Neidhardt and Umbarger 1996), made of

rRNAs and tRNAs that are, respectively, 86 and 14%

(Karpinets et al. 2006). Furthermore, unlike eukaryotic

mRNAs, bacterial mRNA generally lacks poly(A)-tail,

which makes their isolation and analysis complicated (He

et al. 2010).

In this study, we developed, for the first time, a cDNA-

AFLP protocol to study the gene expression profiles of

Lact. rhamnosus strain isolated from Parmigiano Reggiano

cheese. The novel transcriptomic approach was applied to

evaluate the changes in gene expression profile during the

growth of Lact. rhamnosus either in a rich medium

(MRS) or in a cheese-like medium, a medium based on

grated Parmigiano Reggiano ripened cheese, characterized

by the absence of milk sugars and rich in peptides, amino

acids and NaCl (Neviani et al. 2009).

Materials and methods

Bacterial strains and media

Lactobacillus rhamnosus 1473, previously isolated from

20-month ripened Parmigiano Reggiano (PR) cheese,

belonging to biotype no. 1 (Bove et al. 2011) was used in

this study. The strain, isolated on MRS agar (Oxoid Ltd,

Basingstoke, UK), was maintained as stock culture at

)80�C in MRS broth supplemented with 15% glycerol

(w ⁄ v). For RNA extraction, it was first propagated twice,

with a 2% inoculum in 10 ml of MRS broth at 30�C for

24 h; then, the culture was anaerobically (Gas Generating

kit; Oxoid Ltd) propagated twice at 30�C, until the top of

logarithmic phase was reached, in two different media:

MRS broth (De Man et al. 1960) and a cheese-like broth

called cheese broth (CB). CB was prepared partially

cDNA-AFLP for Lactobacillus rhamnosus C.G. Bove et al.

856 Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology

ª 2011 The Authors

modifying the protocol described by Neviani et al. (2009).

Fine-grain grated 20-month ripened PR cheese was dis-

solved (120 g l)1) in 0Æ07 mol l)1 sodium citrate pH 7Æ5(Sigma-Aldrich Co., St Louis, MO, USA). The suspension

was incubated at 42�C for 50 min under stirring condi-

tion (150 g); then, it was centrifuged (8000 g for 15 min

at 4�C) and filtered through sterile gauze to retain the

surfaced fat layer. Filtered liquid was acidified to pH 4Æ6by adding 1 mol l)1 HCl, sterilized for 15 min at 121�C

and centrifuged at 8000 g for 15 min to remove unhydro-

lysed proteins (casein and heat-denatured serum pro-

teins); 1 mol l)1 NaOH was added to supernatant until

pH 6Æ1 was reached. Freeze-dried cell lysate (protein con-

centration of 4 mg ml)1 determined using Bradford pro-

tein assay) of Lactobacillus helveticus PR775, previously

isolated from PR cheese (University of Parma collection),

was added at a final concentration corresponding to

8 log CFU ml)1. The volume of this freeze-dried cell

lysate added was 5 ml in 100 ml of CB (final protein

concentration of 0Æ2 mg ml)1). The resulting medium

was sterilized by filtration on 0Æ22-lm membrane filters

(Millipore Corporation, Bedford, MA, USA). For the

preparation of freeze-dried cell lysate, Lact. helveticus

PR775 was maintained as stock culture at )80�C in MRS

broth supplemented with 15% glycerol (w ⁄ v). It was first

propagated twice, with a 2% inoculum, in MRS broth at

30�C for 24 h; then, the culture was propagated in MRS

at 30�C until the stationary phase was reached. Harvested

cells were washed in 0Æ05 mol l)1 Tris ⁄ HCl pH 7Æ5,

centrifuged (15000 g for 15 min at 4�C) and frozen at

)80�C or directly resuspended in 0Æ05 mol l)1 Tris ⁄ HCl

pH 7Æ5. Cells were disrupted using a Branson model B15

Sonifier by 3 cycles of sonication (1 min each). After

pelleting of unbroken cells (15 000 g for 15 min at 4�C),

the supernatant was maintained as stock at )20�C.

RNA extraction and cDNA synthesis

The growth of Lact. rhamnosus 1473 at 30�C in both MRS

and CB was monitored by optical density (OD) measure-

ment (Jasco, Tokyo, Japan) until the top of logarithmic

phase was reached. About 109 cells at the top of logarith-

mic phase were harvested, and total RNA was stabilized

using RNeasy Protect Bacteria Mini kit (Qiagen, Valencia,

CA, USA) and isolated using RNeasy Mini Kit (Qiagen).

The protocol included the removal of DNA from the sam-

ples by DNAseI using the RNase-Free DNase Set (Qiagen).

Total RNA was resuspended in 50 ll of RNAse-free water

and checked on 0Æ8% agarose gel (Sigma-Aldrich). The

concentration and purity of RNA were assessed by OD

measurement at 260 and 280 nm (Jasco). Prior to cDNA

synthesis, total RNA was enriched in mRNA and then the

mRNA was polyadenylated (Fig. 1).

Using MicrobExpress bacterial mRNA enrichment kit

(Ambion, Applied Biosystems, Monza, Italy), 10 lg of

total RNA from three independent biological replicates

was enriched in mRNA by removing the 16S and 23S

ribosomal RNAs (rRNA). The enriched mRNA was puri-

fied from 5S RNA using the MEGAclear Purification kit

(Ambion) and quantified by spectrophotometer (Jasco).

To add a poly(A) tail to RNA transcripts, we used the

Poly(A) Tailing kit (Ambion). The poly-A RNA (2 lg)

3′5′

(A)n5′

biotinilated5′

PolyA 3′

5′ PolyA 3′

Total RNA

Capture oligo

Oligo MagBead

rRNA molecule

mRNA enrichmentEnriched mRNA

Poly(A) tail added

cDNA synthesis

First restriction digestion

3′ end capturing Streptavidin bead

Second restriction digestion

Adapter ligation

Selective PCR amplificationPrimer 1

Primer 2

(T)n oligo(dT)

Adaptator Adaptator

ge-cDNA-AFLP ce-cDNA-AFLP

Figure 1 Schematic overview of the cDNA-AFLP procedure with 3¢end capture applied to RNA isolated from Lactobacillus rhamnosus

1473. Total RNA isolated from Lact. rhamnosus 1473 is enriched for

mRNA by subtractive hybridization as indicated by the manufacturer

(MicrobExpress; Ambion, Huntington, UK). Poly(A) tail is added to

RNA transcripts using the Poly(A) Tailing kit (Ambion, Applied Biosys-

tems). cDNA is synthesized by reverse transcription using a biotinylat-

ed oligodT, followed by second-strand synthesis. Digestion with first

restriction enzyme captures with the aid of streptavidin-coated

magnetic beads followed by the digestion with the second restriction

enzyme, the removal of the 3¢ fragments and ligation of adaptors.

Selective amplification and separation of the fragments on a polyacryl-

amide gel and capillary electrophoresis and visualization.

C.G. Bove et al. cDNA-AFLP for Lactobacillus rhamnosus

ª 2011 The Authors

Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology 857

was used in the synthesis of double-stranded cDNA

according to the protocol reported by Vuylsteke et al.

(2007) except for the following modifications.

For first-strand complementary DNA synthesis, 1 ll of

100 lmol l)1 oligo dT25bio primer (Eurofins MWG

Operon, Ebersberg, Germany) was added to 20 ll of RNA

sample, and the mixture was heated to 65�C for 5 min

and quick-chilled on ice. The sample was collected by

brief centrifugation, and 4 ll of 5X buffer, 2 ll of DTT

(0Æ1 mol l)1) and 1 ll of dNTPs mix (10 mmol l)1) (In-

vitrogen Life Technologies, Carlsbad, CA, USA) were

added. The sample was kept at 42�C for 2 min to equili-

brate the temperature and then was added with 2 ll of

SuperScript� II RT (200 U ll)1) (Invitrogen Life Tech-

nologies) and incubated at 42�C for 2 h. For second-

strand cDNA synthesis, 32 ll of 5X second-strand buffer,

3 ll of dNTPs mix (10 mmol l)1), 6 ll of DTT

(0Æ1 mol l)1), 1Æ5 ll of E. coli ligase (10 U ll)1), 0Æ2 ll of

ribonuclease H (10 U ll)1), 5 ll of DNA polymerase I

(10 U ll)1) (Invitrogen Life Technologies) and 72Æ3 ll of

H2O were mixed, and all the mixture was added to 40 ll

of first-strand cDNA product. After an incubation of 2 h

at 16�C, the double-stranded cDNA was purified and

concentrated in 30 ll of elution buffer NE using the

NucleoSpin Extract II kits (Macherey-Nagel, Duren,

Germany). The cDNA was checked on 0Æ8% agarose gel

(Sigma-Aldrich) and quantified by spectrophotometer

(Jasco).

Preparation of template for cDNA-AFLP analysis

The template for cDNA-AFLP analysis was prepared

making some modifications to the procedure reported by

Vuylsteke et al. (2007) (Fig. 1). This protocol describes

the generation of cDNA-AFLP fragments detected using

infrared dye (IRD) detection technology and the Odyssey

Infrared Imaging System. A first reaction mixture con-

taining 1 ll of EcoRI (10 U ll)1), 4 ll of NEBuffer 4,

0Æ2 ll of BSA 100X (New England Biolabs, Ipswich, MA,

USA) and 14Æ8 ll of free-nuclease water was added to

20 ll of double-stranded cDNA (about 500 ng). This

reaction mixture was incubated for 2 h at 30�C. Subse-

quently, the mixture was added to 40 ll of solution con-

taining 10 ll of Dynabeads M-280 streptavidin

(Invitrogen Life Technologies) and 40 ll of 2X STEX buf-

fer (40 ml NaCl 5 mol l)1, 2 ml Tris–HCl 1 mol l)1 pH

8Æ0, 400 ll EDTA 0Æ5 mol l)1 pH 8Æ0 and water up to

100 ml). This reaction mixture was incubated for 30 min

at room temperature. This step allows the immobilization

of biotinylated 3¢-terminal cDNA fragments on strepta-

vidin-coated Dynabeads. The Dynabeads were collected,

washed and resuspended in 100 ll of 1X STEX buffer.

Subsequently, a second reaction mixture containing 1 ll

of MseI (10 U ll)1), 4 ll of NEBuffer 4, 0Æ1 ll of BSA

100X (New England Biolabs) and 4Æ9 ll of free-nuclease

water was added. This mixture was incubated for 2 h at

30�C under stirring condition. The supernatant (40 ll)

containing the digested cDNA fragment was added to

10 ll of ligation mixture. This consists in a total volume

of 10 ll containing 1 ll EcoRI adapter (Vuylsteke

et al. 2007) (5 pmol ll)1) (Eurofins MWG Operon), 1 ll

MseI adapter (50 pmol ll)1) (Vuylsteke et al. 2007)

(Eurofins MWG Operon), 1 ll NEBuffer 4 (New England

Biolabs), 1 ll T4 DNA ligase (1 U ll)1), 1 ll ATP

(10 mmol l)1), 0Æ1 ll of BSA 100X (Invitrogen Life Tech-

nologies) and 4Æ9 ll of free-nuclease water. This reaction

mixture was incubated for 3 h at 30�C. Subsequently, the

mixture was diluted to 100 ll with 0Æ1 mol l)1 TE buffer

and stored at )20�C.

Nonselective PCR

The ‘nonselective’ primers EcoRI-0 and MseI-0 (Eurofins

MWG Operon) were used for the pre-amplification of the

diluted template. Each pre-amplification mixture con-

tained 5 ll of the diluted template previously described,

1Æ5 ll of unlabelled MseI-0 primer (10 lmol l)1), 1Æ5 ll

of labelled IRD700EcoRI-0 primer (10 lmol l)1), 0Æ2 ll

of AmpliTaq (5 U ll)1) (Applied Biosystem-Pe Corpora-

tion, Foster City, CA, USA), 5 ll 10X PCR buffer, 5 ll

MgCl2 (25 mmol l)1), 2 ll dNTPs (5 mmol l)1) and

29Æ8 ll of free-nuclease water. The reaction was subjected

to the following PCR conditions: initial denaturation at

94�C for 2 min, 25 cycles consisted of 30-s denaturation

at 94�C, 1-min annealing at 56�C, 1-min extension at

72�C and 1 cycle of final extension for 10 min at 72�C.

All amplifications were performed in a Mastercycler Ep

Gradient S (Eppendorf, Hamburg, Germany). The pre-

amplification product was diluted 600 times with TE

0Æ1 mol l)1 buffer and stored at )20�C.

Selective PCR and capillary electrophoresis (ce-cDNA-

AFLP)

The ‘selective’ primers 5¢FAM-EcoRI-N (N stands for A

or T or AC), labelled with 5¢-carboxy fluorescein (FAM),

and unlabelled MseI-AC ⁄ T (Table 1) (Eurofins MWG

Operon) were used for selective amplification of the

diluted pre-amplification product. PCR mixture con-

tained 5 ll of diluted pre-amplification product, 0Æ8 ll

of labelled 5¢IRD700EcoRI-N primer (1 lmol l)1), 3 ll

of unlabelled MseI-AC ⁄ T primer (2 lmol l)1), 0Æ2 ll of

AmpliTaq (5 U ll)1) (Applied Biosystem-Pe Corpora-

tion), 2 ll of 10X PCR buffer, 2Æ4 ll of MgCl2(25 mmol l)1), 0Æ8 ll of dNTPs (5 mmol l)1) (Invitrogen

Life Technologies) and 5Æ8 ll of free-nuclease water.

cDNA-AFLP for Lactobacillus rhamnosus C.G. Bove et al.

858 Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology

ª 2011 The Authors

The thermocycling conditions were designed as fol-

lows: initial denaturation at 94�C for 2 min, 13 cycles

consisting of 10-s denaturation at 94�C, 30-s annealing

at 65�C (reduced each cycle by 0Æ7�C), 1-min extension

at 72�C, 23 cycles consisting of 10-s denaturation at

94�C, 30-s annealing at 56�C, 1-min extension at 72�C

(extended 1 s per cycle) and 1 cycle of final extension

for 10 min at 72�C. All amplifications were performed

in a Mastercycler Ep Gradient S (Eppendorf). The

cDNA-AFLP analysis of the amplified products was per-

formed making some modification to the procedure

reported by Lazzi et al. (2009). The amplified products

from selective amplification were added to 1 ll of Gene-

Scan-500 (ROX) size standard (Applied Biosystem-Pe

Corporation) and 24 ll of deionized formamide. The mix-

ture was heated for 5 min and cooled for 10 min on ice.

Samples were loaded and run on the ABI Prism 310

(Applied Biosystem-Pe Corporation) according to the

AFLP Microbial Fingerprinting protocol (Applied Bio-

system-Pe Corporation) and analysed using Genemapper

Analysis Software (Applied Biosystem-Pe Corporation)

according to the manufacturer’s instructions. The data for

each run were saved as an individual GeneScan file and

displayed as an electropherogram. A threshold of 100 RFU

(relative fluorescent unit) was considered in scoring to

highlight only sharp and easily distinguishable peaks; all

signals under this value were treated as background and

not scored.

Selective PCR and gel electrophoresis (ge-cDNA-AFLP)

The ‘selective’ primers 5¢IRD700EcoRI-N, labelled with a

infrared dye (IRDye� 700 phosphoramidite), and unla-

belled MseI-AC ⁄ T (Table 1) (Eurofins MWG Operon)

were used for selective amplification of the diluted pre-

amplification product. PCR mixture contained 5 ll of

diluted pre-amplification product, 0Æ8 ll of labelled

5¢IRD700EcoRI-N primer (1 lmol l)1), 3 ll of unlabelled

MseI-AC ⁄ T primer (2 lmol l)1), 0Æ2 ll of AmpliTaq

(5 U ll)1) (Applied Biosystem-Pe Corporation), 2 ll of

10X PCR buffer, 2Æ4 ll of MgCl2 (25 mmol l)1), 0Æ8 ll of

dNTPs (5 mmol l)1) (Invitrogen Life Technologies) and

5Æ8 ll of free-nuclease water.

The thermocycling conditions were the same as

described in ‘Selective PCR and capillary electrophoresis’

section. Electrophoresis was carried out through a 6%

Long Range gel as reported by Vuylsteke et al. (2007).

Amplification products (6 ll) obtained from selective

amplification were mixed with 3 ll of IR2 Stop Solution

Red (LI-COR Biosciences, Lincoln, NE, USA), denatured

at 95�C for 3 min and then cooled on ice for 10 min.

The 50- to 700-bp IRDye700 Sizing Standard (LI-COR

Biosciences) was denatured at 95�C for 2 min and cooled

on ice for 10 min. Samples were loaded onto the pre-

running gel (30 min at 40 mA), and 50- to 700-bp

IRDye700 Sizing Standard (LI-COR Biosciences) was

loaded every 6 lane for normalization purposes. The gel

was run for 4Æ5 h at 20 mA, and after electrophoresis,

Odissey (LI-COR Biosciences) at 700 nm was used for

gel scanning.

Results

In this study, the application of cDNA-AFLP methodol-

ogy for studying the adaptation of Lact. rhamnosus to

cheese environment is reported. Lact. rhamnosus was

grown in CB and MRS media. The top of logarithmic

phase, corresponding to the beginning of stationary

phase, was reached after 20 h by cells growing in MRS

and after 40 h by cells growing in CB (data not shown).

Two techniques (Fig. 1), capillary electrophoresis

cDNA-AFLP (ce-cDNA-AFLP) and gel electrophoresis

cDNA-AFLP (ge-cDNA-AFLP), have been applied to

Table 1 Sequences of the primers and adaptors used for cDNA-AFLP

Primers

and

adaptors Initials Sequences (5¢–3¢) Applications

EcoRI

adaptor

EcoRI Forw CTCGTAGACTGCGTACC Ligation

EcoRI

adaptor

EcoRI Rev AATTGGTACGCAGTCTAC Ligation

MseI

adaptor

MseI Forw GACGATGAGTCCTGAG’ Ligation

MseI

adaptor

MseI Rev TACTCAGGACTCAT Ligation

EcoRI-0 EcoRI-0 GACTGCGTACCAATTC Nonselective

PCR

EcoRI-A 5¢FAM-EcoRI-A GACTGCGTACCAATTCA Selective

PCR

EcoRI-A 5¢IRD-EcoRI-A GACTGCGTACCAATTCA Selective

PCR

EcoRI-T 5¢FAM-EcoRI-T GACTGCGTACCAATTCT’ Selective

PCR

EcoRI-T 5¢IRD-EcoRI-T GACTGCGTACCAATTCT Selective

PCR

EcoRI-AC 5¢FAM-EcoRI-AC GACTGCGTACCAATTCAC Selective

PCR

EcoRI-AC 5¢IRD-EcoRI-AC GACTGCGTACCAATTCAC Selective

PCR

MseI-0 MseI-0 GATGAGTCCTGAGTAA Nonselective

PCR

MseI-T MseI-T GATGAGTCCTGAGTAAT Selective

PCR

MseI-AC MseI-AC GATGAGTCCTGAGTAAAC Selective

PCR

cDNA-AFLP, cDNA-amplified fragment length polymorphism.

C.G. Bove et al. cDNA-AFLP for Lactobacillus rhamnosus

ª 2011 The Authors

Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology 859

obtain and compare the Lact. rhamnosus genome

expressed at maximum growing level in CB and in MRS.

In the first instance, ce-cDNA-AFLP was carried out to

monitor transcriptional changes related to the growth in

two different media. Subsequently, ge-cDNA-AFLP

allowed us to confirm the different expression profiles in

MRS and CB and to obtain fragments that could be

extracted and identified. Principal changes to the original

cDNA-AFLP procedure (Vuylsteke et al. 2007), consisting

in the isolation of mRNA from total RNA and in the

addition of a 3¢ poly(A) tail to mRNA, were made to

develop a novel protocol. This allowed to use an oligodT

biotinylated in the retrotranscription reaction. In this

way, it was possible to retrotranscribe all mRNA using a

unique primer and, subsequently, to capture only one

cDNA fragment deriving from each mRNA species during

the reactions of enzymatic digestion and immobilization

of biotinylated 3¢-terminal cDNA fragments on streptavi-

din-coated Dynabeads.

Double-stranded cDNA synthesized as previously

described from Lact. rhamnosus grown in CB and MRS

was used as template for AFLP analysis. Three combina-

tions of fluorescently labelled primers (Table 2) were

chosen to obtain 50- to 500-bp length fragments. The

selectivity of the primer pairs was determined by one or

two additional nucleotides to increase the stringency of

the PCR amplification step and the sensitivity of the anal-

ysis. In this way, simpler expression profiles were

obtained and analysed on the basis of the pres-

ence ⁄ absence of transcripts. Thanks to the addiction of

selective nucleotides to the primers, low abundance tran-

scripts were also detected as already shown by Breyne

et al. 2003. The different selective primer combinations

generated profiles characterized both by a different num-

ber of fragments and by fragments with different molecu-

lar weight.

In Table 2, the number of fragments as resulting from

capillary and gel electrophoresis is reported. A significant

higher number of fragments are obtained when perform-

ing PCR amplification with the use of the FAM fluores-

cently labelled primer in ce-cDNA-AFLP protocol

(Fig. 2). The fragments generated by ce-cDNA-AFLP,

visualized as peaks in the electropherograms after selective

amplification, ranged from 15 (Fig. 2b1) to 35 (Fig. 2c2)

depending upon primer combinations and growth condi-

tions (Fig. 2).

The highest number of peaks in ce-cDNA-AFLP and

fragments in ge-cDNA-AFLP was produced both in MRS

and in CB using EcoRI-AC ⁄ MseI-T primer combination

(Table 2; Figs 2c1, 2c2 and 3, lanes 5 and 6). Neverthe-

less, the primer pairs EcoRI-A ⁄ MseI-AC and EcoRI-T ⁄MseI-AC originated the higher number of different frag-

ments between the strain grown in MRS and in CB, both

for gel electrophoresis and for capillary electrophoresis

(Table 2). The number of peaks detected in CB through

capillary electrophoresis was always higher than in MRS,

and on the contrary, the number of fragments detected in

CB through gel electrophoresis was lower than in MRS

using EcoRI-AC ⁄ MseI-T (Table 2).

Overall, the use of three pairs of primers allowed

detecting 64 genes expressed in MRS and 97 in CB by

ce-cDNA-AFLP technique. Comparing profiles by gene

mapper software, we found that nine transcripts were

expressed (peak present) by Lact. rhamnosus only when

grown in MRS, while 42 transcripts were expressed by

Lact. rhamnosus only when grown in CB. As a whole, the

profiles of the strain grown in CB were more complex

(Fig. 3).

Discussion

Adaptation of bacteria to their environment involves

many metabolic and physiological changes, and for this

reason, it is considered a very interesting research field.

Lact. rhamnosus species, among lactobacilli, demonstrated

to be able to adapt to very unfavourable environments

such as the ripening of PR cheese and to grow better than

other species becoming a dominant species.

In this study, a novel cDNA-AFLP protocol was

applied to study the profiles of gene expression of one

Lact. rhamnosus strain isolated from 20-month ripened

Parmigiano Reggiano cheese. This is the first research on

Lact. rhamnosus isolated from cheese and represents one

of the few concerning bacterial transcriptomic analysis

Table 2 Primer combinations used in cDNA-AFLP and number of fragments obtained

Primer combination

ce-cDNA-AFLP peaks ge-cDNA-AFLP fragments

MRS CB MRS CB

EcoRI-A ⁄ MseI-AC 16 (Fig. 2a1) 32 (Fig. 2a2) 16 (Fig. 3 lane 1) 24 (Fig. 3 lane 2)

EcoRI-T ⁄ MseI-AC 15 (Fig. 2b1) 29 (Fig. 2b2) 11 (Fig. 3 lane 3) 16 (Fig. 3 lane 4)

EcoRI-AC ⁄ MseI-T 33 (Fig. 2c1) 36 (Fig. 2c2) 33 (Fig. 3 lane 5) 27 (Fig. 3 lane 6)

Total peaks 64 97 60 67

CB, Cheese-like medium, cDNA-AFLP, cDNA-amplified fragment length polymorphism; MRS, rich medium.

cDNA-AFLP for Lactobacillus rhamnosus C.G. Bove et al.

860 Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology

ª 2011 The Authors

towards cDNA-AFLP approaches. In our opinion,

the methodological approach presented, despite its

complexity, could open important perspectives for gene

expression profile studies in bacteria.

The cDNA-AFLP technique is already considered a

valid and alternative approach to microarray for studying

the whole-genome transcription (Bachem et al. 1996;

Reijans et al. 2003; Decorosi et al. 2005). Moreover, the

cDNA-AFLP method reported in our study represents an

innovative approach to study bacteria because it allows

generating unique transcript tags from reverse-transcribed

messenger RNA using restriction enzymes and immobili-

zation of biotinylated 3¢-terminal cDNA fragments on

streptavidin-coated Dynabeads.

The ‘one-gene-one-tag’ approach has been previously

applied only to eukaryotic organism (Vuylsteke et al.

2007), and thanks to protocol modifications made in this

study (Fig. 1), it has been extended to bacteria for the

first time.

The difficulties of working with bacterial mRNA were

bypassed by the subtractive hybridization by rRNA probe

(MicrobExpress bacterial mRNA enrichment kit) that was

found to be most efficient in enriching total RNA in

mRNA and that has already been successfully used in a

recent study on microbiota metatranscriptomic (Booijink

et al. 2010).

Moreover, because the selective cDNA synthesis of

mRNAs based on reverse transcription of the 3¢-poly(A)

tail is not applicable to prokaryotic messengers, a poly(A)

tails to mRNA prior to cDNA synthesis have been added

to obtain a unique transcript for each mRNA (‘one-gene-

one-tag’). The transcript profiles obtained have been

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0

80 120 160 200 240 280 320 360 400 440 480 520

80 120 160 200 240 280 320 360 400 440 480 520

80 120 160 200 240 280 320 360 400 440

80 120 160 200 240 280 320 360 400 440

80 120 160 200 240 280 320 360 400

80 120 160 200 240 280 320 360 400

200

400

600

800

1000

1200

1400

1600

1800

2000

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0

200

400

600

800

1000

1200

1400

1600

1800

2000

0

200

400

600

800

1000

1200

1400

1600

1800

2000(a1)

(a2)

(b1)

(b2)

(c1)

(c2)

Figure 2 Electropherograms of Lactobacillus rhamnosus 1473 cDNA-AFLP profiles using different primer combinations. EcoRI-A ⁄ MseI-AC in MRS

(a1) and CB (a2), EcoRI-T ⁄ MseI-AC in MRS (b1) and CB (b2), EcoRI-AC ⁄ MseI-T in MRS (c1) and CB (c2).

C.G. Bove et al. cDNA-AFLP for Lactobacillus rhamnosus

ª 2011 The Authors

Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology 861

detected using capillary electrophoresis (ce-cDNA-AFLP)

and gel electrophoresis (ge-cDNA-AFLP).

The two applied detection techniques appear to have

different performances: the former (ce-cDNA-AFLP) is

more sensitive and can evidence a greater number of

genes and the latter (ge-cDNA-AFLP) is necessary to

identify them.

The sensitivity of these two techniques is different

because the capacity of gene separation using the gel is

lower than that obtained using capillary electrophoresis.

Furthermore, the lower number of amplification products

generated from cDNA templates in the ge-cDNA-AFLP

methods may have been due to the presence of tran-

scripts with the same or similar molecular weight that

lead to a unique band in the gel. For this reason, the

total number of gel fragments could be lower than the

total number of genes. Thus, fragment detection and

separation through the gel will be improved to optimize

the correspondence between transcripts obtained by two

techniques.

Considering the number of genes checked to date, the

cDNA-AFLP approach resulted able to show that

Lact. rhamnosus modifies the expression of several genes

when cultivated in cheese-like conditions (CB) compared

with growth under in vitro optimal conditions (MRS).

These media have different composition; in fact, MRS is a

standard laboratory complete medium containing a com-

plex nitrogen source consisting of peptone (protein

hydrolysate) and extracts (from meat and yeast), while

CB is a medium based on grated PR ripened cheese, rich

in peptides, amino acids and NaCl but devoid of milk

sugars (Neviani et al. 2009). Although Lact. rhamnosus

was observed to grow better in MRS, it resulted able to

grow as well in the poorer CB. The larger number of

amplification products generated in ce-cDNA-AFLP from

CB-derived template, with respect to those generated

from MRS-derived template, determined a more com-

plex profiles when Lact. rhamnosus grown in cheese like-

environment.

The transcriptomic profiles resulted to be affected by

the growth medium probably because it induces the

activation of different metabolic pathways in the cell to

generate energy and to respond to the environmental

changes. Different studies demonstrated that the adapta-

tion to dairy niches is associated either with a metabolic

simplification as loss of genes involved in carbohydrate,

amino acid and cofactor metabolisms or with an increase

in the expression of genes involved in peptide hydrolysis

and amino acid catabolism (Bolotin et al. 2004; Hols et al.

2005; van de Guchte et al. 2006; Callanan et al. 2008).

In this regard, further analysis will have to be focused

on gel extraction followed by the identification of

expressed genes. This could be interesting to deepen the

knowledge of the basal metabolism through the identifica-

tion of constitutive genes, as well as to analyse how cells

react to the dairy environment through the identification

of differently expressed genes in the two culture media.

50 bp

100 bp

145 bp

200 bp

204 bp

255 bp

300 bp

350 bp

364 bp

400 bp

700 bp

M 1 2 3 4 5 6 M

Figure 3 cDNA-AFLP fingerprint obtained with two different primer

combinations. M, 50- to 700-bp IRDye700 Sizing Standard; lanes 1

and 2, cDNA-AFLP fingerprinting of Lactobacillus rhamnosus cultured

in MRS and CB, respectively, using EcoRI-A ⁄ MseI-AC primer combina-

tion; lanes 3 and 4, cDNA-AFLP fingerprinting of Lact. rhamnosus

cultured in MRS and CB, respectively, using EcoRI-T ⁄ MseI-AC primer

combination; lanes 5 and 6, cDNA-AFLP fingerprinting of

Lact. rhamnosus cultured in MRS and CB using EcoRI-AC ⁄ MseI-T

primer combination.

cDNA-AFLP for Lactobacillus rhamnosus C.G. Bove et al.

862 Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology

ª 2011 The Authors

Moreover, the expected results could lead to under-

stand whether Lact. rhamnosus physiological mechanisms

of adaptation to cheese-like substrates, which are

fundamental for survival in cheese during long ripening,

could be involved both in ripening processes and in the

definition of organoleptic characteristics of Parmigiano

Reggiano cheese.

Acknowledgements

The authors are grateful to Dr Elisa Sgarbi and Antonietta

Cirasolo for technical assistance.

References

Bachem, C.W., van der Hoeven, R.S., de Bruijn, S.M., Vreug-

denhil, D., Zabeau, M. and Visser, R.G. (1996) Visualiza-

tion of differential gene expression using a novel method

of RNA fingerprinting based on AFLP: analysis of gene

expression during potato tuber development. Plant 9, 745–

753.

Banks, J.M. and Williams, A.G. (2004) The role of nonstarter

lactic acid bacteria in Cheddar cheese ripening. Int J Dairy

Technol 57, 145–152.

Bensch, S. and Akesson, M. (2005) Ten years of AFLP in

ecology and evolution: why so few animals? Mol Ecol 14,

2899–2914.

Beresford, T. and Williams, A. (2004) The microbiology of

cheese ripening. In Cheese: Chemistry, Physics and Micro-

biology ed Fox, P.F., McSweeney, P.L.H., Cogan, T.M. and

Guinee, T.P. pp. 287–318. the Netherlands: Elsevier,

Amsterdam.

Beresford, T.P., Fitzsimons, N.A., Brennan, N.L. and Cogan,

T.M. (2001) Recent advances in cheese microbiology. Int

Dairy J 11, 259–274.

Bolotin, A., Quinquis, B., Renault, P., Sorokin, A., Ehrlich,

S.D., Kulakauskas, S., Lapidus, A., Goltsman, E. et al.

(2004) Complete sequence and comparative genome analy-

sis of the dairy bacterium Streptococcus thermophilus. Nat

Biotechnol 22, 1554–1558.

Booijink, C.C.G.M., Boekhorst, J., Zoetendal, E.G., Smidt, H.,

Kleerebezem, M. and de Vos, W.M.V. (2010) Metatran-

scriptome analysis of the human fecal microbiota reveals

subject-specific expression profiles, with genes encoding

proteins involved in carbohydrate metabolism being

dominantly expressed. Appl Environ Microbiol 76,

5533–5540.

Bove, C.G., De Dea Lindner, J., Lazzi, C., Gatti, M. and

Neviani, E. (2011) Evaluation of genetic polymorphism

among Lactobacillus rhamnosus nonstarter Parmigiano

Reggiano cheese strains. Int J Food Microbiol 144, 569–572.

Breyne, P. and Zabeau, M. (2001) Genome-wide expression

analysis of plant cell cycle modulated genes. Curr Opin

Plant Biol 4, 136–142.

Breyne, P., Dreesen, R., Cannoot, B., Rombaut, D., Vandepo-

ele, K., Rombauts, S., Vanderhaeghen, R., Inze, D. et al.

(2003) Quantitative cDNA-AFLP analysis for genome-wide

expression studies. Mol Gen Genomics 269, 173–179.

Broadbent, J.R. and Steele, J.L. (2005) Cheese flavour and the

genomics of lactic acid bacteria. ASM News 71, 121–125.

Callanan, M., Kaleta, P., O’Callaghan, J., O’Sullivan, O., Jor-

dan, K., McAuliffe, O., Sangrador-Vegas, A., Slattery, L.

et al. (2008) Genome sequence of Lactobacillus helveticus,

an organism distinguished by selective gene loss and inser-

tion sequence element expansion. J Bacteriol 190, 727–735.

De Dea Lindner, J., Bernini, V., De Lorentiis, A., Pecorari, A.,

Neviani, E. and Gatti, M. (2008) Parmigiano Reggiano

cheese: evolution of cultivable and total lactic microflora

and peptidase activities during manufacture and ripening.

Dairy Sci Technol 88, 511–523.

De Man, J.C., Rogosa, M. and Sharpe, M.E. (1960) A medium

for the cultivation of lactobacilli. J Appl Bacteriol 23, 130–

135.

Decorosi, F., Vitia, C., Mengoni, A., Bazzicalupo, M. and

Giovannetti, L. (2005) Improvement of the cDNA-AFLP

method using fluorescent primers for transcription analysis

in bacteria. J Microbiol Methods 63, 211–215.

Dellagi, A., Birch, P.R., Heilbronn, J., Lyon, G.D. and Toth,

I.K. (2000) cDNA-AFLP analysis of differential gene

expression in the prokaryotic plant pathogen Erwinia

carotovora. Microbiology 146, 165–171.

Ekkapongpisit, M., Wannatung, T., Susantad, T.,

Triwitayakorn, K. and Smith, D.R. (2007) cDNA-AFLP

analysis of differential gene expression in human hepatoma

cells (HepG2) upon dengue virus infection. J Med Virol

79, 552–561.

Gatti, M., De Dea Lindner, J., De Lorentiis, A., Bottari, B.,

Santarelli, M., Bernini, V. and Neviani, E. (2008)

Dynamics of whole and lysed bacterial cells during

Parmigiano-Reggiano cheese production and ripening.

Appl Environ Microbiol 74, 6161–6167.

van de Guchte, M., Penaud, S., Grimaldi, C., Barbe, V.,

Bryson, K., Nicolas, P., Robert, C., Oztas, S. et al. (2006)

The complete genome sequence of Lactobacillus bulgaricus

reveals extensive and ongoing reductive evolution. Proc

Natl Acad Sci U S A 103, 9274–9279.

He, S., Wurtzel, O., Singh, K., Froula, J.I., Yilmaz, S., Tringe,

S.G., Wang, Z., Chen, F. et al. (2010) Validation of two

ribosomal RNA removal methods for microbial metatrans-

criptomics. Nat Methods 7, 807–813.

Hols, P., Hancy, F., Fontaine, L., Grossiord, B., Prozzi, D.,

Leblond-Bourget, N., Decaris, B., Bolotin, A. et al. (2005)

New insights in the molecular biology and physiology of

Streptococcus thermophilus revealed by comparative

genomics. FEMS Microbiol Rev 29, 435–463.

Kadota, K., Araki, R., Nakai, Y. and Abe, M. (2007) GOGOT:

a method for the identification of differentially expressed

fragments from cDNA-AFLP data. Algorithms Mol Biol 2,

5.

C.G. Bove et al. cDNA-AFLP for Lactobacillus rhamnosus

ª 2011 The Authors

Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology 863

Karpinets, T.V., Greenwood, D.J., Sams, C.E. and Ammons,

J.T. (2006) RNA: protein ratio of the unicellular organism

as a characteristic of phosphorous and nitrogen stoichio-

metry and of the cellular requirement of ribosomes for

protein synthesis. BMC Biol 4, 30.

Kieronczyk, A., Skeie, S., Olsen, K. and Langsrud, T. (2001)

Metabolism of amino acids by resting cells of non-starter

lactobacilli in relation to flavour development in cheese.

Int Dairy Journal 11, 217–224.

Koskenniemi, K., Koponen, J., Kankainen, M., Savijoki, K.,

Tynkkynen, S., de Vos, W.M., Kalkkinen, N. and

Varmanen, P. (2009) Proteome analysis of Lactobacillus

rhamnosus GG using 2-D DIGE and mass spectrometry

shows differential protein production in laboratory and

industrial-type growth media. J Proteome Res 8, 4993–

5007.

Lazzi, C., Bove, C.G., Sgarbi, E., Gatti, M., La Gioia, F.,

Torriani, S. and Neviani, E. (2009) Application of AFLP

fingerprint analysis for studying the biodiversity of

Streptococcus thermophilus. J Microbiol Methods 79, 48–

54.

Levterova, V., Panaiotov, S., Brankova, N. and Tankova, K.

(2010) Typing of genetic markers involved in stress

response by fluorescent cDNA-amplified fragment length

polymorphism technique. Mol Biotechnol 45, 34–38.

Liu, S.-Q., Holland, R. and Crow, V.L. (2003) The potential of

dairy lactic acid bacteria to metabolise amino acids via

non-transaminating reactions and endogenous transamina-

tion. Int J Food Microbiol 86, 257–269.

Neidhardt, F.C. and Umbarger, H.F. (1996) Chemical compo-

sition of Escherichia coli. In Escherichia coli and Salmonella

typhimurium, Cellular and Molecular Biology, 2nd edn. ed.

Neidhart, F.C., Ingraham, J.L., Low, K.B., Magasanik, B.,

Schaechter, M. and Umbarger, H.E. pp. 13–17. Washington,

DC: ASM Press.

Neveu, C., Charvet, C., Fauvin, A., Cortet, J., Castagnone-

Sereno, P. and Cabaret, J. (2007) Identification of

levamisole resistance markers in the parasitic nematode

Haemonchus contortus using a cDNA-AFLP approach.

Parasitology 134, 1105–1110.

Neviani, E., De Dea Lindner, J., Bernini, V. and Gatti, M.

(2009) Recovery and differentiation of long ripened cheese

microflora through a new cheese-based cultural medium.

Food Microbiol 26, 240–245.

Noel, L., Thieme, F., Nennstiel, D. and Bonas, U. (2001)

cDNA-AFLP analysis unravels a genome-wide hrpG-

regulon in the plant pathogen Xanthomonas campestris pv.

vesicatoria. Mol Microbiol 41, 1271–1281.

Reijans, M., Lascaris, R., Groeneger, A.O., Wittenberg, A.,

Wesselink, E., van Oeveren, J., de Wit, E., Boorsma, A.

et al. (2003) Quantitative comparison of cDNA-AFLP,

microarrays, and GeneChip expression data in

Saccharomyces cerevisiae. Genomics 82, 606–618.

Settanni, L. and Moschetti, G. (2010) Non-starter lactic acid

bacteria used to improve cheese quality and provide health

benefits. Food Microbiol 27, 691–697.

Succi, M., Tremonte, P., Reale, A., Sorrentino, E., Grazia, L.,

Pacifico, S. and Coppola, R. (2005) Bile salt and acid

tolerance of Lactobacillus rhamnosus strains isolated from

Parmigiano Reggiano cheese. FEMS Microbiol Lett 244,

129–137.

Vuylsteke, M., Peleman, J.D. and van Eijk, M.J.T. (2007)

AFLP-based transcript profiling (cDNAAFLP) for

genome-wide expression analysis. Nat Protoc, 2, 1399–

1413.

Weiberg, A., Pohler, D., Morgenstern, B. and Karlovsky, P.

(2008) Improved coverage of cDNA-AFLP by sequential

digestion of immobilized cDNA. BMC Genomics 9, 480.

Xiaohu, X., Heping, L. and Chaorong, T. (2009) A silver-

staining cDNA-AFLP protocol suitable for transcript

profiling in the latex of Hevea brasiliensis (para rubber

tree). Mol Biotechnol 42, 91–99.

Yvon, M. and Rijnen, L. (2001) Cheese flavour formation by

amino acid catabolism. Int Dairy J 11, 185–201.

cDNA-AFLP for Lactobacillus rhamnosus C.G. Bove et al.

864 Journal of Applied Microbiology 111, 855–864 ª 2011 The Society for Applied Microbiology

ª 2011 The Authors